Tracking Surface Oil

Ian R. MacDonaldFlorida State University

White et al. 2016 Oceanography 29 76-87

Overview

• Oil spill methods—where does remote sensing fit in?

• A flow-chart of remote sensing applications

• Lessons from the Gulf—DWH and beyond

• The problem of thickness

• At the cutting edge—bring on the drones

Cycle of Remote Sensing Applications Applied to a Unique Spill

Detection

• Genuine spill?

• Where?

• Ongoing?

Evaluation

• Quantity?

• Rate?

• Actionable?

Tactical Response

• Deploy assets?

• Movement?

• Containment?

Injury & Damage Assessment

• Distribution

• Resource intersection

• Impact

• Lessons learned

1504 m

Characterize Surface Oil Distributions Using Optical Remote Sensing

– Thickness estimates obtained by processing AVIRIS images from a limited survey of DwH suggested a median thickness of 70 µm for emulsified oil (Sun, Hu et al. 2015).

– The DwH SAR data set was subsequently used to quantify “thin” (1 µm) and “thick” (70 µm) oil.

AVIRIS-derived oil thickness (left) and Red-Green-Blue

true color maps (from Clark et al., 2010)

MacDonald et al. (in review) Remote Sensing Assessment of Surface Oil Transport and Fate during Spills in the Gulf of Mexico, BOEM Report

Priority Oil—Based on Ensemble Sensor Suite

Characterize Surface Oil Distributions Using Synthetic Aperture Radar

Satellite SAR image processing was performed using TCNNA algorithm (Garcia-Pineda, Zimmer et al. 2009, Garcia-Pineda, MacDonald et al. 2010)

TCNNA is an FSU developed algorithm which employs satellite and meteorological variables, and textural analysis to extract oil features from RADAR images.

Each SAR pixel is classified as feature or non-feature and to reproduce, pixel by pixel, classifications an expert human analyst would make evaluating whether a given group of pixels belongs to a feature.

For WAMOST, additional routines were developed to output multiple values corresponding to apparent oil thickness categories.

MacDonald et al. (in review) Remote Sensing Assessment of Surface Oil Transport and Fate during Spills in the Gulf of Mexico, BOEM Report

Oil Emulsion Detection Algorithm

– A publication for WAMOST (Garcia-Pineda, MacDonald et al. 2013) describes the development of the Oil Emulsion Detection Algorithm.

– The Deepwater Horizon spill generated large areas of rainbow sheen and smaller regions of emulsified oil.

– Comparison of aerial photos and surface samples with SAR images found that emulsion produced a radar signature with intermediated intensity between unoiled water and floating sheen.

– This signature was used to segment areas of oil emulsion in 60 SAR images, which were applied to a time-series analysis of the spill.

MacDonald et al. (in review) Remote Sensing Assessment of Surface Oil Transport and Fate during Spills in the Gulf of Mexico, BOEM Report

Characterize Surface Oil Distributions Using Optical Remote Sensing

Inflection point:

~ 70 m

Inflection point:

~ 1 m

95% line

Statistics of AVIRIS-derived oil thickness shows two inflection values around 1 and 70 m, which were used in SAR analysis to derive oil volume time-series (MacDonald et al. 2015)

Surface volume: 12-h best estimate from SAR

Average Volume (m3/km2)

Sea-Trial of Subsea Dispersant Use• Responders were

stressed by HC fumes when working over the DWH site.

• A application of Corexit 95 into the subsea plume was tested on morning 10 May.

• Result was a temporary reduction of “thick” surface oil.

Lehr, W., S. Bristol and A. Possolo (2010). "Oil Budget Calculator Deepwater Horizon Technical Documentation." The Federal Interagency Solutions Group

MacDonald et al. 2015 doi:10.1002/2015JC011062 Fig. 6

Time Series of DWH OilOil-Covered Water—all thicknesses

6 May↓

23 May↓

18 June↓

18 July↓

Time Series of DWH OilOil-Covered Water—Thick Oil (~70 µm)

6 May↓

23 May↓

18 June↓

18 July↓

Time Series of DWH OilDaily SAR Volume of Surface Oil

6 May↓

23 May↓

18 June↓

18 July↓

MacDonald et al. 2015 doi:10.1002/2015JC011062 Fig. 7

Saildronewww.saildrone.com

• Autonomous surface sailor.• Solar power—extended

endurance• Intelligent steering• Near real-time reporting from

array of sensors• Challenging conditions in the

GoM—low wind, many obsticles

Daneshgar et al. Hindcast modelling for the persistence of floating oil slicks released from natural seeps. Remote Sensing Environ. (In review)

Ohmsett Drone Remote Sensing 2016

Take-home ideas

• Remote sensing plays an essential role for detection and assessment of marine oil spills.

– Expansion of application could include routine satellite collections for monitoring at-risk areas.

• Satellite assets have fixed orbital schedules, which limit their tactical applications during response efforts.

– Relating oil thickness & chemical state to satellite images is an area where research and development is needed.

• Drone technology is highly promising, but needs field trials to make operational.